AI News

Machine-learning clustering of close-in exoplanet populations: links to pebble accretion (arxiv.org)

TacCoRL: Integrating Tactile Feedback into VLA via Simulation (arxiv.org)

Last-Iterate Convergence of Optimistic Multiplicative Weight Update (arxiv.org)

Deterministic Policy Gradient for Learning Equilibrium in Time-Inconsistent Control Problems (arxiv.org)

SPEA2$^+$: Improved Density Estimation in SPEA2 with Provable Runtime Guarantees (arxiv.org)

Neural FOXP2 -- Language Specific Neuron Steering for Targeted Language Improvement in LLMs (arxiv.org)

TAHOE: Text-to-SQL with Automated Hint Optimization from Experience (arxiv.org)

System Report for CCL25-Eval Task 5: New Dataset and LoRA-Fine-Tuned Qwen2.5 (arxiv.org)

DIRECT: When and Where Should You Allocate Test-Time Compute in Embodied Planners? (arxiv.org)

Critic Architecture Matters: Dual vs. Unified Critics for Humanoid Loco-Manipulation (arxiv.org)

REACH: Interpretability-Driven Feature Identification and Architecture Compression for Multi-Channel Vehicular Channel Estimation (arxiv.org)

Modelling magnetic material properties with uncertainty-aware neural networks (arxiv.org)

Seeing Below the Limit of Detection: A Censored-Poisson Bayesian Latent-Growth Change-Point Detector (the Span Detector) for Serial ctDNA in HR+/HER2- Metastatic Breast Cancer (arxiv.org)

GraspLLM: Towards Zero-Shot Generalization on Text-Attributed Graphs with LLMs (arxiv.org)

Reroute, Don't Remove: Recoverable Visual Token Routing for Vision-Language Models (arxiv.org)

Autoregressive Direct Preference Optimization (arxiv.org)

Offline Diffusion Policy for Multi-User Delay-Constrained Scheduling (arxiv.org)

MLaGA: Multimodal Large Language and Graph Assistant (arxiv.org)

Continual Learning with Support Boundary Experience Blending (arxiv.org)

NARRAS: Edge-Triggered Distributed Inference for CSI-Based Localization in Vehicular IoT Networks (arxiv.org)

Corpus Augmentation for Sign Language Translation via LLM-Guided Video Stitching (arxiv.org)

Neuro-Relational Programs: Unifying Queries and Neural Computation over Structured Data (arxiv.org)

Categorical Robustness Assessment for Machine Learning based Network Intrusion Detection Systems (arxiv.org)

DAM-VLA: Decoupled Asynchronous Multimodal Vision Language Action model (arxiv.org)

Beyond Dark Knowledge: Mixup-Based Distillation for Reliable Predictions (arxiv.org)

Sustainability assessment using multimodal AI agents (arxiv.org)

Resource-Aware LLM Reasoning for Mobile Edge General Intelligence (arxiv.org)

An XAI View on Explainable ASP: Methods, Systems, and Perspectives (arxiv.org)

Making Models Unmergeable via Scaling-Sensitive Loss Landscape (arxiv.org)

Quantum Occam Learning: Sample-Supported Expressibility for Circuit-Based Quantum Learning (arxiv.org)

MLT-Dedup: Efficient Large-Scale Online Video Deduplication via Multi-Level Representations and Spatial-Temporal Matching (arxiv.org)

Finding Sparse Subnetworks in One Training Cycle via Progressive Magnitude-Based Pruning (arxiv.org)

PianoKontext: Expressive Performance Rendering from Deadpan Context (arxiv.org)

Learning What to Say to Your VLA: Mostly Harmless Vision Language Action Model Steering (arxiv.org)

Measuring Semantic Progress in Multi-turn Dialogue via Information Gain (arxiv.org)

Adjoint Method versus Physics-Informed Neural Networks in PDE-Constrained Inverse Problems (arxiv.org)

Diffusion-based Cumulative Adversarial Purification for Vision Language Models (arxiv.org)

Cross-Layer Discrete Concept Discovery for Interpreting Language Models (arxiv.org)

MARIC: Multi-Agent Reasoning for Image Classification (arxiv.org)

UniIntervene: Agentic Intervention for Efficient Real-World Reinforcement Learning (arxiv.org)

Context-Driven Incremental Compression for Multi-Turn Dialogue Generation (arxiv.org)

Learning Patterns and Abstractions from Perceptual Sequences (arxiv.org)

AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin (arxiv.org)

Persistent Homology as a Theory of Emergent Structure (arxiv.org)

Tensor Methods: A Unified and Interpretable Approach for Material Design (arxiv.org)

Energy Use of AI Inference, Efficiency Pathways, and Test-Time Scaling (arxiv.org)

Time-multiplexed layer reuse for physical neural networks (arxiv.org)

CaReTS: A Multi-Task Framework Unifying Classification and Regression for Time Series Forecasting (arxiv.org)

Visualizing LLM Latent Space Geometry Through Dimensionality Reduction (arxiv.org)

MobileFineTuner: A Mobile-Native Framework for On-Device LLM Fine-Tuning in Real-World Embedded AI Applications (arxiv.org)